Multi-resolution decomposition is applied to tower data sets to study the cospectral gap scale, which is the time scale that separates turbulent and mesoscale fluxes of heat, moisture and momentum between the atmosphere and the surface. We develop a simple model for the gap time scale as a function of height above ground and bulk atmospheric stability to provide guidance in choosing the local mean time scale for making improved eddy correlation flux calculations from tower data. Similarity relationships are improved by using the adjustable averaging time scale model to calculate turbulent fluxes compared to using a constant time scale.